Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Boltz Structure Prediction

Predict protein complex and protein–ligand structures with open-source Boltz-1/Boltz-2 and validate designed binders.

Data & AnalyticsAdvanced★ 178⑂ 32AI score 8/10Last updated: Jul 1, 2026

What it does

  • Runs Boltz-1/Boltz-2, an open-weights biomolecular structure predictor, for protein–protein complexes, protein–ligand complexes and binder validation.
  • Gives two execution paths: Modal cloud GPU (modal run modal_boltz.py) or a local install (pip install boltz → boltz predict).
  • Documents key parameters (--recycling_steps, --sampling_steps, --use_msa_server) with defaults and safe ranges.
  • Explains the output layout (CIF model, confidence.json, PAE matrix), pass thresholds (pTM > 0.7, ipTM > 0.5, pLDDT > 0.7) and a Biopython snippet to convert CIF → PDB.
  • Includes an error table for CUDA OOM, missing iptm (single-chain FASTA), missing weights and non-standard residues.
  • Adds a decision tree for choosing Boltz vs Chai vs ColabFold vs ESMFold, plus runtime/cost estimates per campaign size.

Who it's for

  • Protein designers who need structural validation of de novo binders
  • Teams that want open weights instead of AF2's licensing constraints
  • Computational biology engineers with a 24GB+ GPU or serverless GPU credits (Modal)

Examples

  1. "Predict the target–binder complexes in complex.fasta and drop anything with ipTM < 0.5."
  2. "How long and how much to fold 500 complexes on Modal?" → ~2–3 hours, roughly $35 from the performance table.
  3. "I hit CUDA out of memory" → switch to MSA-free mode, reduce sampling_steps, or move to an A100-80GB.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/BioTender-max/awesome-bio-agent-skills/HEAD/skills/adaptyv/boltz/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/adaptyv/boltz folder from the GitHub repo BioTender-max/awesome-bio-agent-skills into my ~/.claude/skills/boltz/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/BioTender-max/awesome-bio-agent-skills.git /tmp/abas && mkdir -p ~/.claude/skills/boltz && cp -r /tmp/abas/skills/adaptyv/boltz/* ~/.claude/skills/boltz/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal and clone the repo: git clone https://github.com/BioTender-max/awesome-bio-agent-skills.git
  2. Create the skill folder: mkdir -p ~/.claude/skills/boltz
  3. Copy the skill files: cp -r awesome-bio-agent-skills/skills/adaptyv/boltz/* ~/.claude/skills/boltz/
  4. Restart Claude Code and ask something like "use boltz to predict this complex" to confirm the skill loads.
  5. Prepare the runtime: Python 3.10+, CUDA 12+, 24GB+ VRAM; run pip install boltz then boltz download to fetch model weights.
  6. For cloud runs, create a Modal account, run pip install modal && modal setup, then use modal_boltz.py from the biomodals repo.
  7. Make sure your FASTA has two or more chains (>protein_A, >protein_B) so interface confidence (ipTM) is reported.
View source on GitHub ↗License: NOASSERTION